Install a Helm Chart on GKE

Install Kanban on GKE#

We’ll use our umbrella chart to install an application with only one command. Yes, that’s right. With only one command we can install an application on a cloud!

Here is a summary of the most important commands that can help us manage a cluster.

And here is the command to install the Kanban application, which might take a couple of minutes:

The output will be as follows:

Release "kanban" does not exist. Installing it now.
W0107 06:53:14.496052   22612 warnings.go:70] Autopilot set default resource requests for Deployment kanban/kanban-frontend, as resource requests were not specified. See http://g.co/gke/autopilot-defaults.  
W0107 06:53:14.497283   22612 warnings.go:70] Autopilot set default resource requests for Deployment kanban/kanban-backend, as resource requests were not specified. See http://g.co/gke/autopilot-defaults.   
W0107 06:53:14.609664   22612 warnings.go:70] Autopilot increased resource requests for StatefulSet kanban/postgres to meet requirements. See http://g.co/gke/autopilot-resources.
NAME: kanban
LAST DEPLOYED: Fri Jan  7 06:53:13 2022
NAMESPACE: kanban
STATUS: deployed
REVISION: 1
Thank you for running Kanban application!


1. Establish port-forwarding:

   > kubectl port-forward svc/kanban-frontend 8080:8080 --namespace kanban

2. Enter the page http://localhost:8080/

If you encounter, Error: Kubernetes cluster unreachable: Get "http://localhost:8080/version": dial tcp 127.0.0.1:8080: connect: connection refused make sure that you execute a following command:

Let’s try it out in the playground below:

Verify installation#

And that’s it! All the hard work from the previous lesson is finally paying off!

Locally#

Now, let’s check if everything is working. First, let’s get all the Kubernetes resources from the kanban namespace:

The output will be as follows:

NAME                                   READY   STATUS    RESTARTS   AGE
pod/kanban-backend-69fcf4979b-5hbdk    1/1     Running   1          3m52s
pod/kanban-backend-69fcf4979b-zxfpr    1/1     Running   0          3m52s
pod/kanban-frontend-55cf65f8f8-rqcmk   1/1     Running   0          3m52s
pod/postgres-0                         1/1     Running   0          3m52s

NAME                        TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)    AGE
service/kanban-backend      ClusterIP   10.10.129.122   <none>        8080/TCP   3m52s
service/kanban-frontend     ClusterIP   10.10.131.199   <none>        8080/TCP   3m52s
service/postgres            ClusterIP   10.10.128.77    <none>        5432/TCP   3m52s
service/postgres-headless   ClusterIP   None            <none>        5432/TCP   3m52s

NAME                              READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/kanban-backend    2/2     2            2           3m52s
deployment.apps/kanban-frontend   1/1     1            1           3m52s

NAME                                         DESIRED   CURRENT   READY   AGE
replicaset.apps/kanban-backend-69fcf4979b    2         2         2       3m53s
replicaset.apps/kanban-frontend-55cf65f8f8   1         1         1       3m53s

NAME                        READY   AGE
statefulset.apps/postgres   1/1     3m53s

We can drill down to get more details, e.g., for kanban-backend deployment, as shown below:

The output will be as follows:

Name:                   kanban-backend
Namespace:              kanban
CreationTimestamp:      Fri, 07 Jan 2022 13:56:49 +0100
Labels:                 app=kanban-backend
                        app.kubernetes.io/managed-by=Helm
                        app.kubernetes.io/name=kanban-backend
                        app.kubernetes.io/version=kanban-1
                        group=backend
Annotations:            autopilot.gke.io/resource-adjustment:
                          {"input":{"containers":[{"name":"kanban-backend"}]},"output":{"containers":[{"limits":{"cpu":"500m","ephemeral-storage":"1Gi","memory":"2G...
                        deployment.kubernetes.io/revision: 1
                        meta.helm.sh/release-name: kanban
                        meta.helm.sh/release-namespace: kanban
Selector:               app=kanban-backend
Replicas:               2 desired | 2 updated | 2 total | 2 available | 0 unavailable
StrategyType:           RollingUpdate
MinReadySeconds:        0
RollingUpdateStrategy:  25% max unavailable, 25% max surge
Pod Template:
  Labels:  app=kanban-backend
           app.kubernetes.io/name=kanban-backend
           app.kubernetes.io/version=kanban-1
           group=backend
  Containers:
   kanban-backend:
    Image:      wkrzywiec/kanban-app:helm-course
    Port:       80/TCP
    Host Port:  0/TCP
    Limits:
      cpu:                500m
      ephemeral-storage:  1Gi
      memory:             2Gi
    Requests:
      cpu:                500m
      ephemeral-storage:  1Gi
      memory:             2Gi
    Environment:
      DB_SERVER:          postgres
      POSTGRES_DB:        kanban
      POSTGRES_PASSWORD:  kanban
      POSTGRES_USER:      kanban
    Mounts:               <none>
  Volumes:                <none>
Conditions:
  Type           Status  Reason
  ----           ------  ------
  Progressing    True    NewReplicaSetAvailable
  Available      True    MinimumReplicasAvailable
OldReplicaSets:  <none>
NewReplicaSet:   kanban-backend-69fcf4979b (2/2 replicas created)
Events:
  Type    Reason             Age    From                   Message
  ----    ------             ----   ----                   -------
  Normal  ScalingReplicaSet  4m49s  deployment-controller  Scaled up replica set kanban-backend-69fcf4979b to 2

If we want to test an application by ourselves, the only thing to do is to set up port-forwarding, as it was done previously (shown below):

Forwarding from 127.0.0.1:8080 -> 80
Forwarding from [::1]:8080 -> 80

However, on the platform, execute the following command:

So locally, if we enter http://localhost:8443 in a web browser or, on the platform, we click on the URL provided in the widget, we should be able to see Kanban’s main page, as shown below:

On Google Cloud#

We should be sure by now that the application is working. But before moving to the next lesson, we would recommend another check in the Google Cloud dashboard. To do that, go to https://console.cloud.google.com, and from the left menu select “Kubernetes Engine” (as shown below).

We should see a list of clusters with only a single entry. On the left, a new menu appears. We can use it to get a more detailed insight into our clusters, similar to the Kubernetes Dashboard. Here is a view of the “Workloads” page which lists all the Deployments and StatefulSets.

We can select one of them, e.g., kanban-backend, and we’ll be directed to the detailed page. We can get more insight about the Deployment, including all the Pods that are part of it, as shown below.

If we select one of them, we are then redirected to the Pod’s detailed page

And if we click “Logs” we can get access to all the logs that are produced by our application, as shown below: